brice, r. (2003). newnes guide to digital tv (2nd ed.)

68 Practical digital audio interface.... 98 More about digital filtering and signal processing..... Preface to the first editionNewnes Guide to Digital Television is written for those wh

An imprint of Elsevier Science

Linacre House, Jordan Hill, Oxford OX2 8DP

200 Wheeler Road, Burlington, MA 01803

First published 2000

Second edition 2003

Copyright # 2000, 2003, Richard Brice All rights reserved

No part of this publication may be reproduced in any

material form (including photocopying or storing in

any medium by electronic means and whether or not

transiently or incidentally to some other use of this

publication) without the written permission of the

copyright holder except in accordance with the provisions

of the Copyright, Designs and Patents Act 1988

or under the terms of a licence issued by the Copyright

Licensing Agency Ltd, 90 Tottenham Court Road, London,

England W1T 4LP Applications for the copyright holder’s

written permission to reproduce any part of this publication

should be addressed to the publishers

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN 0 7506 5721 9

For information on all Newnes publications, visit

our website at www.newnespress.com

Data manipulation by David Gregson Associates, Beccles, Suffolk

Printed and bound in Great Britain by Antony Rowe Ltd., Chippenham, Wiltshire

Preface to the first edition xv

1 Introduction 1

Digital television 1

Why digital? 1

More channels 3

Wide- screen pictures 3

Cinema sound 4

Associated services 4

Conditional access 6

Transmission techniques 6

Receiver technology 6

The future 7

2 Foundations of television 8

A brief history of television 8

The introduction of colour 9

The physics of light 9

Physiology of the eye 10

Psychology of vision ^ colour perception 12

Metamerism ^ the great colour swindle 13

Persistence of vision 14

The physics of sound 14

Fourier 16

Transients 16

Physiology of the ear 17

Psychology of hearing 18

Masking 19

Temporalmasking 20

Film and television 21

Television 22

Television signals 24

H sync and V sync 24

Colour television 26

NTSC and PAL colour systems 27

SECAMcolour system 31

Shadowmask tube 32

Vestigial sideband modulation 33

Audio for television 34

NICAM728 digital stereo sound 35

Recording television signals 35

Colour under 36

Audio tracks 38

Timecode 38

Longitudinal timecode 38

Vertical interval timecode (VITC) 40

PAL and NTSC 40

User bits 40

TeletextTM 40

Analogue high definition television ( HDTV) 41

MAC 43

PALplus 43

1125/ 60 and 1250/ 50 HDTV systems 44

1250/50 European HDTV 44

625- line television wide screen signalling 44

TheWSS signal 44

Data structure 44

Display formats 46

Telecine and pulldown 46

3 Digital video and audio coding 50

Digital fundamentals 50

Sampling theory and conversion 51

Theory 51

Themechanismof sampling 53

Aliasing 54

Quantization 54

Digital-to-analogue conversion 55

Jitter 56

Aperture effect 56

Dither 56

Digital video interfaces 57

Clock signal 61

Filter templates 61

Parallel digital interface 62

Serial digital interface 63

HDTV serial interface 65

Digital audio interfaces 66

AES/EBU or IEC958 type 1 interface 66

SPDIF or IEC958 type 2 interface 67

Data 68

Practical digital audio interface 70

TOSlink optical interface 70

Unbalanced (75 ohm) AES interface 71

Serial multi- channel audio digital interface ( MADI) 72

Data format 73

Scrambling and synchronization 76

Electrical format 77

Fibre optic format 77

Embedded audio in video interface 77

Error detection and handling 80

EDH codeword generation 81

EDH flags 83

4 Digital signal processing 85

Digital manipulation 86

Digital filtering 86

Digital image processing 88

Point operations 88

Window operations 89

Transforming between time and frequency domains 93

Fourier transform 93

Phase 95

Windowing 96

2- D Fourier transforms 98

More about digital filtering and signal processing 99

Convolution 100

Impulse response 100

FIR and IIR filters 101

Design of digital filters 102

Frequency response 103

Derivation of band-pass and high-pass filters 105

Designing an IIR filter 106

IIR filter design example 107

High-pass filter example 109

Digital frequency domain analysis ^ the z- transform 110

Problems with digital signal processing 110

5 Video data compression 112

Entropy, redundancy and artefacts 112

Lossless compression 113

De-correlation 114

Lossless DPCM and lossy DPCM 116

Frame differences and motion compensation 117

Fourier transform- based methods of compression 119

Transform coding 119

A practicalmix 123

JPEG 125

Motion JPEG ( MJPEG) 127

MPEG 127

Levels and profiles 128

Main profile atmain level 129

Main level at 4 : 2 : 2 profile (ML@4 : 2 : 2P) 129

Frames or fields 129

MPEG coding 131

Mosquito noise 135

MPEG coding hardware 135

Statisticalmultiplexing 136

DV, DVCAM and DVCPRO 137

6 Audio data compression 139

Compression based on logarithmic representation 139

NICAM 140

Psychoacoustic masking systems 140

MPEG layer I compression (PASC) 141

MPEG layer III 143

Dolby AC- 3 143

7 Digital audio production 145

Digital line- up levels and metering 145

The VUmeter 146

The PPMmeter 147

Opto-electronic level indication 148

Standard operating levels and line- up tones 149

Digital line-up 149

Switching and combining audio signals 150

Digital audio consoles 151

Soundmixer architecture 151

Mixer automation 152

Digital tape machines 153

Digital two-track recording 153

Digitalmulti-tracks 154

Digital audio workstations 155

Audio file formats 156

WAV files 156

AU files 157

AIFF and AIFC 157

MPEG 157

VOC 158

Raw PCMdata 158

Surround- sound formats 158

Dolby Surround 158

Dolby digital (AC-3) 161

Rematrixing 161

Dynamic range compression 161

MPEG-II extension tomulti-channel audio 162

Pro logic compatibility 162

IEC 61937 interface 162

Dynamic range compression 163

Multilingual support 163

EditingMPEG layer II audio 163

8 Digital video production 164

Swi4tching and combining video signals 164

Digital video effects 166

What is a video transition? 166

The cut 167

The dissolve 167

The fade 169

Wipes 169

Split-screens 170

Keys 170

Posterize 171

Chroma-key 172

Off- line editing 174

Computer video standards 175

Vector and bitmap graphics ^ whats the difference? 177

Graphic file formats 178

Windows Bitmap (.BMP) 179

PCX 179

TARGA 179

GIF 179

JPEG 180

Computer generated images (CGI) and animation 180

Types of animation 181

Software 182

2D systems 182

Paint-system functions 182

Compositing 188

Morphing and warping 189

Rotorscoping 190

3D graphics and animation 191

Matrices 192

Imaging 194

Light 195

Ray tracing 197

Hard disk technology 198

Winchester hard disk drive technology 199

Hard drive interface standards 200

IDE drives 200

SCSI 201

Fibre channel 201

Firewire 201

RAID 202

RAID 1 (mirroring) 203

RAID 2 (bit striping with error correction) 203

RAID 3 (bit striping with parity) 203

RAID 4 (striping with fixed parity) 204

RAID 5 (striping with striped parity) 204

Media server 204

Open media framework 205

Virtual sets 205

The master control room 205

Automation 206

Editing and switching of MPEG-II bitstreams 208

The ATLANTIC Project 209

Mole 209

9 The MPEG multiplex 210

A packetized interface 210

Deriving the MPEG-II multiplex 211

The PES packet format 211

Transport stream 213

Packet synchronization 213

Packet identification 213

Programassociation tables and programmap tables 213

Error handling 214

The adaptation header 214

Synchronization and timing signals 214

Systemand program clock references 215

Presentation timestamps 215

Splicing bitstreams 216

Conditional access table 216

DVB service information 216

Conditional access 217

SimulCrypt andMultiCrypt 217

Channel coding 218

Randomization (scrambling) 218

Reed-Solomon encoding 219

Convolutional interleaving 220

Standard electrical interfaces for the MPEG- II transport stream 221

Synchronous parallel interface 221

Synchronous serial interface 222

The asynchronous serial interface 223

10 Broadcasting digital video 225

Digital modulation 225

Quadrature amplitude modulation 225

Modulation for satellite and cable systems 229

Establishing reference phase 230

Convolutional or Viterbi coding 230

Terrestrial transmission ^ DVB-T (COFDM) and US ATSC ( 8- VSB) systems 231

Coded orthogonal frequency divisionmultiplexing (COFDM) 232

Practical COFDM 233

Adding a guard period to OFDMmodulation 233

The advantages of COFDM 234

8-VSBmodulation 235

Hierarchical modulation 237

Interoperability 238

Interoperability with ATM 238

ATMcell and transport packet structures 239

11 Consumer digital technology 240

Receiver technology 240

Current set-top box design 242

Circuit descriptions 243

Set- top box ^ modern trends 253

Digital tuner 253

Incorporation of hard disk drives (PVR) 257

D- VHS 258

DVD 259

Track structure 259

Data rates and picture formats 260

Audio 261

Control 262

Regional codes 262

The DVD player 263

CPSA (content protection system architecture) 264

Analogue copy protection 264

Copy generationmanagement system (CGMS) 264

Content scrambling system 265

DVD Recordable ( DVD- R) 266

General servicing issues 267

Static and safety 267

Golden rules 267

Equipment 268

DVD faults 268

PSU faults 269

12 The future 270

Leaning forward and leaning back 270

Hypertext and hypermedia 271

HTML documents 271

Anchor tags 272

Images 273

MPEG-IV ^ object-oriented television coding 273

Objects and scenes 274

The language 275

Virtual realitymodelling language (VRML) 275

Practical VRML files 282

MPEG- IV audio 284

Structured audio 285

Structured audio orchestra language 285

Text-to-speech systems 286

Audio scenes 288

MPEG- VII and metadata 289

Index 291

Preface to the second edition

In the four years or so since I started work on the first edition of this book,digital television has changed from being a reality to a common place Myown family is proof of this Whether the children are watching a widechoice of programmes via digital cable in Paris or, more recently, a similarchoice carried by digital satellite in the UK, they have come to expect thebenefits that digital television brings; more channels, better quality,interactivity etc In addition to this ubiquity, there have been some realtechnical developments too The inclusion of a hard drive within the set-top box, speculated on in the first edition, has now become a reality withthe personal video recorder or PVR permitting the tapeless time-shifting ofprogrammes and true video-on-demand (VOD) movie channels

But it is not simply in the broadcast areas of television that we have seenhuge advances in the last few years The adoption of the DV camcorderand the proliferation of digital editing software on personal computerplatforms has spread digital technology to the videographer as well as thebroadcaster Most of all, the DVD and the availability of reasonable pricedwide-screen televisions have changed the public’s perception of thequality boundaries of the experience of watching television This edition,therefore, has sought to cover in more detail these trends and develop-ments and, to that end, you will find herein expanded sections on DVD,the inclusion of sections on DV video compression and how it differs fromMPEG

As a general rule, nothing dates so comprehensively in a technical book

as a chapter titled ‘The future’! However, lest one believe that the pace ofchange makes it impossible for the non-specialist to follow, I am pleased

to say that digital television has undergone several years of consolidationrather than revolution and the final chapter remains as relevant as it waswhen it was originally written

xiii

Once again, I should like to thank those who are mentioned in the Preface

to the First Edition, and add my thanks to Neil Sharpe of MirandaTechnologies Ltd for permission to use the photographs of the MirandaDV-Bridge and Presmaster mixer My family too, deserve to be thankedagain for their forbearance in having a father who writes books instead ofspending more time with them

Richard BriceGreat Coxwell 2002xiv Preface to the second edition

Preface to the first edition

Newnes Guide to Digital Television is written for those who are faced withthe need to comprehend the novel world of digital television technology.Not since the 1960s – and the advent of colour television in Europe – havemanagers, technicians and engineers had to learn so much, so quickly;whether they work in the development laboratory, the studio or in therepair-shop This book aims to cover the important principles that lie atthe heart of the new digital TV services I have tried to convey the broadarchitecture of the various systems and how these offer the functionalitiesthey do By concentrating on important principles, rather than presentingreams of detail, I hope the important ideas presented will ‘stick in themind’ more obstinately than if I had adopted the opposite approach I amalso aware that there exists a new generation of engineers ‘in the wings’,

as it were, to whom the world of digital television will be the onlytelevision they will know For them, I have included a chapter on theimportant foundations of television as they have evolved in the first 60years of this interesting and world-changing technology

Acknowledgements

I think, if we’re honest, most engineers who work in television wouldagree that the reality of transmitted digital television has crept up on us all.Like a lion, it has circled for the last 20 years, but the final pounce hastaken many by surprise In truth, I should probably have started writingNewnes Guide to Digital Television earlier than I did But I don’t think it’sjust retrospective justification to say that I hesitated because, even a shorttime ago, many technologies that are today’s reality were still in thelaboratory and research information was very thin indeed There hastherefore been the need to make up for lost time in the publishing phase Ishould also like to thank Andy Thorne and Chris Middleton of DesignSphere of Fareham in the UK, who were extremely helpful and patient as I

xv

wrestled to understand current set-top box technology; their help wasinvaluable and greatly appreciated It is with their permission that thecircuit diagrams of the digital set-top receiver appear in Chapter 11.Finally, my thanks and apologies to Claire, who has put up with a newhusband cloistered in his study when he should have been repairing thebathroom door!

Richard BriceParis, 1999xvi Preface to the first edition

Why digital?

I like to think of the gradual replacement of analogue systems with digitalalternatives as a slice of ancient history repeating itself When the ancientGreeks – under Alexander the Great – took control of Egypt, the Greeklanguage replaced Ancient Egyptian and the knowledge of how to writeand read hieroglyphs was gradually lost Only in 1799 – after a period of

2000 years – was the key to deciphering this ancient written languagefound following the discovery of the Rosetta stone Why was this knowl-edge lost? Probably because Greek writing was based on a writtenalphabet – a limited number of symbols doing duty for a whole language.Far better, then, than the seven hundred representational signs of AncientEgyptian writing Any analogue system is a representational system – awavy current represents a wavy sound pressure and so on Hieroglyphicelectronics if you like! The handling and processing of continuous time-variable signals (like audio and video waveforms) in digital form has allthe advantages of a precise symbolic code (an alphabet) over an older

1

Figure 1.1 Guide to Digital Television ^ Route map

approximate representational code (hieroglyphs) This is because, oncerepresented by a limited number of abstract symbols, a previouslyundefended signal may be protected by sending special codes, so thatthe digital decoder can work out when errors have occurred For example,

if an analogue television signal is contaminated by impulsive interferencefrom a motorcar ignition, the impulses (in the form of white and blackdots) will appear on the screen This is inevitable, because the analoguetelevision receiver cannot ‘know’ what is wanted modulation and what isnot A digital television can sort the impulsive interference from wantedsignal As television consumers, we therefore expect our digital televisions

to produce better, sharper and less noisy pictures than we have come toexpect from analogue models (Basic digital concepts and techniques arediscussed in Chapter 3; digital signal processing is covered in Chapter 4.)

More channels

So far, so good, but until very recently there was a down side to digitalaudio and video signals, and this was the considerably greater capacity, orbandwidth, demanded by digital storage and transmission systemscompared with their analogue counterparts This led to widespreadpessimism during the 1980s about the possibility of delivering digitaltelevision to the home, and the consequent development of advancedanalogue television systems such as MAC and PALplus However, thedisadvantage of greater bandwidth demands has been overcome byenormous advances in data compression techniques, which make betteruse of smaller bandwidths In a very short period of time these techniqueshave rendered analogue television obsolescent It’s no exaggeration to saythat the technology that underpins digital television is data compression orsource coding techniques, which is why this features heavily in the pagesthat follow An understanding of these techniques is absolutely crucial foranyone technical working in television today Incredibly, data compres-sion techniques have become so good that it’s now possible to put manydigital channels in the bandwidth occupied by one analogue channel;good news for viewers, engineers and technicians alike as more oppor-tunities arise within and without our industry

Wide-screen pictures

The original aspect ratio (the ratio of picture width to height) for the motionpicture industry was 4 : 3 According to historical accounts, this shape wasdecided somewhat arbitrarily by Thomas Edison while working withGeorge Eastman on the first motion picture film stocks The 4 : 3 shapethey worked with became the standard as the motion picture business grew.Today, it is referred to as the ‘Academy Standard’ aspect ratio When the first

experiments with broadcast television occurred in the 1930s, the 4 : 3 ratiowas used because of historical precedent In cinema, 4 : 3 formatted imagespersisted until the early 1950s, at which point Hollywood studios began torelease ‘wide-screen’ movies Today, the two most prevalent film formatsare 1.85 : 1 and 2.35 : 1 The latter is sometimes referred to as ‘Cinemascope’

or ‘Scope’ This presents a problem when viewing wide-screen cinemareleases on a 4 : 3 television In the UK and in America, a technique known as

‘pan and scan’ is used, which involves cropping the picture The alternative,known as ‘letter-boxing’, presents the full cinema picture with black bandsacross the top and bottom of the screen Digital television systems allprovide for a wide-screen format, in order to make viewing film releases(and certain programmes – especially sport) more enjoyable Note thatdigital television services don’t have to be wide-screen, only that thestandards allow for that option Television has decided on an intermediatewide-screen format known as 16 : 9 (1.78 : 1) aspect ratio Figure 1.2illustrates the various film and TV formats displayed on 4 : 3 and 16 : 9 TVsets Broadcasters are expected to produce more and more digital 16 : 9programming Issues affecting studio technicians and engineers are covered

in Chapters 7 and 8

‘Cinema’ sound

To complement the wide-screen cinema experience, digital television alsodelivers ‘cinema sound’; involving, surrounding and bone-rattling! For solong the ‘Cinderella’ of television, and confined to a 5-cm loudspeaker atthe rear of the TV cabinet, sound quality has now become one of thestrongest selling points of a modern television Oddly, it is in the soundcoding domain (and not the picture coding) that the largest differences liebetween the European digital system and the American incarnation TheEuropean DVB project opted to utilize the MPEG sound coding method,whereas the American infrastructure uses the AC-3 system due to DolbyLaboratories For completeness, both of these are described in thechapters that follow; you will see that they possess many more similaritiesthan differences: Each provides for multi-channel sound and for asso-ciated sound services; like simultaneous dialogue in alternate languages.But more channels mean more bandwidth, and that implies compressionwill be necessary in order not to overload our delivery medium This isindeed the case, and audio compression techniques (for both MPEG andAC-3) are fully discussed in Chapter 6

Figure 1.2 Different aspect ratios displayed 4: 3 and 16 : 9

Such a complex environment means not only will viewers need helpnavigating between channels, but the equipment itself will also requiredata on what sort of service it must deliver: In the DTV standards, user-definable fields in the MPEG-II bitstream are used to deliver serviceinformation (SI) to the receiver This information is used by the receiver

to adjust its internal configuration to suit the received service, and can also

be used by the broadcaster or service provider as the basis of an electronicprogramme guide (EPG) – a sort of electronic Radio Times ! There is nolimit to the sophistication of an EPG in the DVB standards; manybroadcasters propose sending this information in the form of HTMLpages to be parsed by an HTML browser incorporated in the set-topbox Both the structure of the MPEG multiplex and the incorporation ofdifferent types of data are covered extensively in Chapter 9

This ‘convergence’ between different digital media is great, but itrequires some degree of standardization of both signals and the interfacesbetween different systems This issue is addressed in the DTV world as thedegree of ‘interoperability’ that a DTV signal possesses as it makes the

‘hops’ from one medium to another These hops must not cause delays orloss of picture and sound quality, as discussed in Chapter 10

Conditional access

Clearly, someone has to pay for all this technology! True to their birth

in the centralist milieu of the 1930s, vast, monolithic public analoguetelevision services were nurtured in an environment of nationallyinstituted levies or taxes; a model that cannot hope to continue in theeclectic, diversified, channel-zapping, competitive world of today For thisreason, all DTV systems include mechanisms for ‘conditional access’,which is seen as vital to the healthy growth of digital TV These issuestoo are covered in the pages that follow

Transmission techniques

Sadly, perhaps, just as national boundaries produced differing analoguesystems, not all digital television signals are exactly alike All current andproposed DTV systems use the global MPEG-II standard for image coding;however, not only is the sound-coding different, as we have seen, but the

RF modulation techniques are different as well, as we shall see in detail inlater chapters

receivers – in incredible numbers! A survey of digital television would bewoefully incomplete without a chapter devoted to receiver and set-topbox technology as well as to digital versatile disc (DVD), which is oustingthe long-treasured VHS machine and bringing digital films into increasingnumbers of homes.

One experience is widespread in the engineering community associatedwith television in all its guises; that of being astonished by the rate ofchange within our industry in a very short period of time Technology thathas remained essentially the same for 30 years is suddenly obsolete, and agreat many technicians and engineers are aware of being caught un-prepared for the changes they see around them I hope that this book willhelp you feel more prepared to meet the challenges of today’s television.But here’s a warning; the technology’s not going to slow down! Today’stelevision is just that – for today The television of next year will bedifferent For this reason I’ve included the last chapter, which outlinessome of the current developments in MPEG coding that will set theagenda of television in the future In this way I hope this book willserve you today and for some years to come

Foundations of television

Of course, digital television didn’t just spring fully formed from theground! Instead it owes much to its analogue precursors No one candoubt digital television represents a revolution in entertainment but, at atechnological level, it is built on the foundations of analogue television.More than this, it inherited many presumptions and constraints from itsanalogue forebears For this reason, an understanding of analoguetelevision techniques is necessary to appreciate this new technology;hence the inclusion of this chapter You will also find here a briefdescription of the psychological principles that underlie the development

of this new television technology

A brief history of television

The world’s first fully electronic broadcast television service was launched

by the BBC in London in November 1936 The system was the result of thevery great work by Schoenberg and his team at the EMI Company Initiallythe EMI system shared the limelight with Baird’s mechanical system, butthe latter offered poor quality by comparison and was quickly dropped infavour of the all-electronic system in February 1937 In the same year,France introduced a 455-line electronic system and Germany and Italyfollowed with a 441-line system Oddly, the United States were some waybehind Europe, the first public television service being inaugurated inNew York in 1939; a 340-line system operating at 30 frames/second Twoyears later, the United States adopted the (still current) 525-line standard.Due to the difficulty of adequate power-supply decoupling, earlytelevision standards relied on locking picture rate to the AC mainsfrequency as this greatly ameliorated the visible effects of hum Hencethe standards schism that exists between systems of American andEuropean origin (the AC mains frequency is 60 Hz in the North Americarather than 50 Hz in Europe) In 1952 the German GERBER system was

8

proposed in order to try to offer some degree of harmonization betweenAmerican and European practice It was argued that this would ease thedesign of standards conversion equipment, and thereby promote thegreater transatlantic exchange of television programmes; as well asenabling European television manufacturers the opportunity to exploitthe more advanced American electronic components To this end, the linefrequency of the GERBER system was chosen to be very close to the 525-line American system but with a frame rate of 50, rather than 60, fields persecond The number of lines was thereby roughly defined by

ð525
60Þ=50 ¼ 630The GERBER system was very gradually adopted throughout Europeduring the 1950s and 1960s

The introduction of colour

Having been a little slow off the mark in launching an electronic TVservice, television in the USA roared ahead with the introduction, in 1953,

of the world’s first commercial colour television service, in which colourinformation is encoded in a high-frequency subcarrier signal Standardized

by the National Television System Committee, this system is known wide as the NTSC system Eight years later in France, Henri de Franceinvented the Sequentiel Couleur a Memoire system (SECAM) which usestwo alternate subcarriers and a delay-line ‘memory store’ AlthoughSECAM requires a very greatly more complicated receiver than NTSC (anot inconsequential consideration in 1961), it has the advantage that thecolour signal can suffer much greater distortion without perceptibleconsequences At about the same time – and benefiting from thetechnology of ultrasonic delay lines developed for SECAM – Dr WalterBruch invented the German PAL system, which is essentially a modifiedNTSC system PAL retains some of the robustness of SECAM, whilstoffering something approaching the colour fidelity of NTSC Colourtelevision was introduced in the UK, France and Germany in 1967, 14years after the introduction of the NTSC system

world-Now let’s look at some of the perceptual and engineering principleswhich underlie television; analogue and digital

The physics of light

When an electromotive force (EMF) causes an electric current to flow in awire, the moving electric charges create a magnetic field around the wire.Correspondingly, a moving magnetic field is capable of creating an EMF in

an electrical circuit These EMFs are in the form of voltages constrained to

dwell inside electric circuits However, they are special cases of electricfields The same observation could be phrased: a moving electric fieldcreates a magnetic field and a moving magnetic field creates an electricfield This affords an insight into a form of energy that can propagatethrough a vacuum (i.e without travelling in a medium) by means of anendless reciprocal exchange of energy shunted backwards and forwardsbetween electric and magnetic fields This is the sort of energy that light is.Because it is itself based on the movement of electricity, it’s no surprisethat this type of energy has to move at the same speed that electricity does– about 300 million metres per second Nor that it should be dubbedelectromagnetic energy, as it propagates or radiates through empty space

in the form of reciprocally oscillating magnetic and electric fields known

as electromagnetic waves

Although the rate at which electromagnetic energy radiates throughspace never changes, the waves of this energy may vary the rate at whichthey exchange electric field for magnetic field and back again Indeed,these cycles vary over an enormous range Because the rate at which theenergy moves is constant and very fast, it’s pretty obvious that, if the rate

of exchange is relatively slow, the distance travelled by the wave tocomplete a whole cycle is relatively vast The distance over which a wave

of electromagnetic energy completes one cycle of its repeating pattern ofexchanging fields is known as the wavelength of the electromagneticenergy It’s fair to say that the range of wavelengths of electromagneticenergy boggle the human mind, for they stretch from cosmic rays withwavelengths of a hundred million millionth of a metre to the energyradiated by AC power lines with wavelengths of a million million metres!For various physical reasons only a relatively small region of this hugerange of energy, which floods from all over the universe and especiallyfrom our sun, arrives on the surface of the earth The range that doesarrive has clearly played an important role in the evolution of life, sincethe tiny segment of the entire diapason is the range to which we areattuned and that we are able to make use of A small part of this smallrange is the range we call light Wavelengths longer than light, extending

to about a millimetre, we experience as heat

Physiology of the eye

The wavelengths the human eye perceives extend only from about

380 nm to about 780 nm, in frequency terms, just over one octave.Visual experience may occur by stimulation other than light waves –pressure on the eyeball, for example; an observation that indicates that theexperience of light is a quality produced by the visual system Put anotherway, there’s nothing special about the range of electromagneticwavelengths 380–780 nm, it’s just that we experience them differently

10 Newnes Guide to Digital TV

from all the others We shall see that colour perception, too, is a function

of the perceptual system and not a physical attribute of electromagneticradiation

Physiologically, the eye is often compared to a camera because theyboth consist of a chamber, open at one end to let in the light, and avariable lens assembly for focusing an image on a light-sensitive surface atthe rear of the chamber In the case of the camera, the light-sensitivematerial is film; in the case of the eye, the retina Figure 2.1 illustrates thehuman eye in cross-section

A rough 2.5-cm sphere, the human eye bulges at the front in the region

of the cornea – a tough membrane that is devoid of a blood supply inorder to maintain good optical properties The cornea acts with the lenswithin the eye to focus an image on the light-sensitive surface at the rear

of the vitreal cavity The eye can accommodate (or focus images atdifferent distances) because the lens is not rigid but soft, and its shapemay be modified by the action of the ciliary muscles These act via thesuspensory ligament to flatten the lens in order to view distant objects, andrelax to view near objects The iris, a circular membrane in front of thelens, is the pigmented part of the eye that we see from the outside The iris

is the eye’s aperture control, and controls the amount of light entering theeye through the opening in the iris known as the pupil

Figure 2.1 The physiology of the eye

The light-sensitive surface at the back of the eye known as the retinahas three main layers:

1 Rods and cones – which are photosensitive cells that convert lightenergy into neural signals;

2 Bipolar cells, which make synaptic connections with the rods andcones;

3 Ganglion cells, which form the optic nerve through which the visualsignals are passed to the vision-processing regions of the brain.Experiments with horseshoe crabs, which possess a visual system notablyamenable to study, have revealed that the light intensity falling upon eachvisual receptor is conveyed to the brain by the rate of nerve firings In ourpresent context, the cells of interest are the 100 million cylindrical rodsand the 6 million more bulbous cones that may be found in one singleretina The cones are only active in daylight vision, and permit us to seeboth achromatic colours (white, black and greys – known as luminanceinformation) and colour The rods function mainly in reduced illumina-tion, and permit us to see only luminance information So it is to the cones,which are largely concentrated in the central region of the retina known asthe fovea, that we must look for the action of colour perception

Psychology of vision ^ colour perception

Sir Isaac Newton discovered that sunlight passing through a prism breaksinto the band of multicoloured light that we now call a spectrum Weperceive seven distinct bands in the spectrum:

red, orange, yellow, green, blue, indigo, violet

We see these bands distinctly because each represents a particular band ofwavelengths The objects we perceive as coloured are perceived thusbecause they too reflect a particular range of wavelengths For instance, adaffodil looks yellow because it reflects predominantly wavelengths in theregion 570 nm We can experience wavelengths of different colourbecause the cones contain three photosensitive chemicals, each of which

is sensitive in three broad areas of the light spectrum It’s easiest to think ofthis in terms of three separate but overlapping photochemical processes; alow-frequency (long wavelength) red process, a medium-frequency greenprocess and a high-frequency blue process (as electronic engineers, youmight prefer to think of this as three shallow-slope band-pass filters!).When light of a particular frequency falls on the retina, the action of thelight reacts selectively with this frequency-discriminating mechanism.When we perceive a red object, we are experiencing a high level ofactivity in our long wavelength (low-frequency) process and low levels inthe other two A blue object stimulates the short wavelength or high-

12 Newnes Guide to Digital TV

frequency process, and so on When we perceive an object with anintermediate colour, say the yellow of the egg yoke, we experience amixture of two chemical process caused by the overlapping nature of each

of the frequency-selective mechanisms In this case, the yellow light fromthe egg causes stimulation in both the long wavelength red process and themedium wavelength green process Because human beings possess threeseparate colour vision processes, we are classified as trichromats Peopleafflicted with colour blindness usually lack one of the three chemicalresponses in the normal eye; they are known a dichromats, although a fewrare individuals are true monochromats What has not yet been discovered,amongst people or other animals, is a more-than-three colour perceptionsystem This is lucky for the engineers who developed colour television!

Metamerism ^ the great colour swindle

The fact that our cones only contain three chemicals is the reason that wemay be fooled into experiencing the whole gamut of colours with thecombination of only three so-called primary colours The televisionprimaries of red, green and blue were chosen because each stimulatesonly one of the photosensitive chemicals found in the cone cells Thegreat colour television swindle (known technically at metamerism) is that

we can, for instance, be duped into believing we are seeing yellow byactivating both the red and green tube elements simultaneously – just aswould a pure yellow source (see Figure 2.2) Similarly, we may behoodwinked into seeing light-blue cyan with the simultaneous activation

of green and blue We can also be made to experience paradoxicalcolours like magenta by combining red and blue, a feat that no pure light

Figure 2.2 Photochemical response of the chemicals in the eye

source could ever do! This last fact demonstrates that our colourperception system effectively ‘wraps-around’, mapping the linear spec-trum of electromagnetic frequencies into a colour circle, or a colour space.

It is in this way that we usually view the science of colour perception; wecan regard all visual sense as taking place within a colour three-space Atelevision studio vectorscope allows us to view colour three-space end-on,

so it looks like a hexagon (Figure 2.3b) Note that each colour appears at adifferent angle, like the numbers on a clock face ‘Hue’ is the term used inimage processing and television to describe a colour’s precise location onthis locus ‘Saturation’ is the term used to describe the amount a purecolour is diluted by white light The dashed axis shown in Figure 2.3a isthe axis of pure luminance The more a particular shade moves towardsthis axis from a position on the boundary of the cube, the more a colour issaid to be de-saturated

Persistence of vision

The human eye exhibits an important property that has great relevance tothe film and video industries This property is known as the persistence ofvision When an image is impressed upon the eye, an instantaneouscessation of the stimulus does not result in a similarly instantaneouscessation of signals within the optic nerve and visual processing centres.Instead, an exponential ‘lag’ takes place, with a relatively long timerequired for total decay Cinema and television have exploited thiseffect for over 100 years

The physics of sound

Sound waves are pressure variations in the physical atmosphere Thesetravel away at about 300 metres per second in the form of waves, which

14 Newnes Guide to Digital TV

Figure 2.3 (a) Colour three-space; (b) Colour three-space viewedend-on as in a TV vectorscope

spread out like ripples on a pond In their journey, these waves collidewith the walls, chairs, tables – whatever – and make them move ever soslightly The waves are thus turned into heat and ‘disappear’ These wavescan also cause the fragile membrane of the eardrum to move Exactly whathappens after that is a subject we’ll look at later in the chapter All thatmatters now is that this movement is experienced as the phenomenon wecall hearing.

It is a demonstrable property of all sound sources that they oscillate: anoboe reed vibrates minutely back and forth when it is blown; the air inside

a flute swells and compresses by an equal and opposite amount as it isplayed; a guitar string twangs back and forth Each vibration is termed acycle The simplest sound is elicited when a tone-producing objectvibrates backwards and forwards, exhibiting what physicists call simpleharmonic motion When an object vibrates in this way it follows the pathtraced out in Figure 2.4; known as a sine wave

Such a pure tone, as illustrated, actually sounds rather dull andcharacterless But we can still vary such a sound in two important ways.First, we can vary the number of cycles of oscillation that take place persecond Musicians refer to this variable as pitch; physicists call itfrequency The frequency variable is referred to in hertz (Hz), meaningthe number of cycles that occur per second Secondly, we can alter itsloudness; this is related to the size, rather than the rapidity, of theoscillation In broad principle, things that oscillate violently produceloud sounds This variable is known as the amplitude of the wave.Unfortunately, it would be pretty boring music that was made up solely

of sine tones despite being able to vary their pitch and loudness Thewaveform of a guitar sound is shown in Figure 2.5 As you can see, theguitar waveform has a fundamental periodicity like the sine wave, butmuch more is going on If we were to play and record the waveform ofother instruments each playing the same pitch note, we would notice asimilar but different pattern; the periodicity would remain the same, butthe extra small, superimposed movements would be different The termFigure 2.4 Sine wave

we use to describe the character of the sound is ‘timbre’, and the timbre of

a sound relates to these extra movements which superimpose themselvesupon the fundamental sinusoidal movement that determines the funda-mental pitch of the musical note Fortunately these extra movements areamenable to analysis too; in fact, in a quite remarkable way

Fourier

In the eighteenth century, J.B Fourier – son of a poor tailor who roseultimately to scientific advisor to Napoleon – showed that any signal thatcan be generated can be alternatively expressed as a sum of sinusoids ofvarious frequencies With this deduction, he gave the world a whole newway of comprehending waveforms Previously only comprehensible as atime-based phenomena, Fourier gave us new eyes to see with Instead ofthinking of waveforms in the time base (or the time domain) as we seethem displayed on an oscilloscope, we may think of them in the frequencybase (or the frequency domain) comprised of the sum of various sinewaves of different amplitudes and phase.1 In time, engineers have given

us the tools to ‘see’ waveforms expressed in the frequency domain too.These are known as spectrum analysers or, eponymously, as Fourieranalysers (see Figure 2.6) The subject of the Fourier transform, whichbestows the ability to translate between these two modes of description, is

so significant in many of the applications considered hereafter that awhole section is devoted to this transform in Chapter 4

Transients

The way a musical note starts is of particular importance in our ability torecognize the instrument on which it is played The more characteristicand sharply defined the beginning of a note, the more rapidly we are able

to determine the instrument from which it is elicited This bias towardtransient information is even evident in spoken English, where we useabout 16 long sounds (known as phonemes) against about 27 short

16 Newnes Guide to Digital TV

Figure 2.5 Guitar waveform

phonemes Consider the transient information in a vocalized list of wordsthat end the same way; coat, boat, dote, throat, note, wrote, tote and votefor instance! Importantly, transients too can be analysed in terms of acombination of sinusoids of differing amplitudes and phases using theFourier integral as described above.

Physiology of the ear

Studies of the physiology of the ear reveal that the process of Fourieranalysis, referred to earlier, is more than a mere mathematical conception.Anatomical and psychophysiological studies have revealed that the earexecutes something very close to a mechanical Fourier analysis on thesounds it collects, and passes a frequency domain representation of thosesounds on to higher neural centres An illustration of the human ear isgiven in Figure 2.7

After first interacting with the auricle or pinna, sound waves travel downthe auditory canal to the eardrum The position of the eardrum marks theboundary between the external ear and the middle ear The middle ear is

an air-filled cavity housing three tiny bones; the hammer, the anvil and thestirrup These three bones communicate the vibrations of the eardrum tothe oval window on the surface of the inner ear Due to the manner inwhich these bones are pivoted, and because the base of the hammer isbroader than the base of the stirrup, there exists a considerable mechan-ical advantage from eardrum to inner ear A tube runs from the base of themiddle ear to the throat; this is known as the Eustachian tube Its action is

to ensure that equal pressure exists on either side of the eardrum, and it isopen when swallowing The inner ear is formed in two sections; thecochlea (the spiral structure which looks like a snail’s shell) and the threesemicircular canals These latter structures are involved with the sense ofbalance and motion

Figure 2.6 A time-domain and frequency domain representation

The stirrup is firmly attached to the membrane that covers the ovalwindow aperture of the cochlea The cochlea is full of fluid and is dividedalong its entire length by the Reissner’s membrane and the basilarmembrane, upon which rests the organ of Corti When the stirrupmoves, it acts like a piston at the oval window and this sets the fluidwithin the cochlea into motion This motion, trapped within the enclosedcochlea, creates a standing wave pattern – and therefore a distortion – inthe basilar membrane Importantly, the mechanical properties of thebasilar membrane change considerably along its length As a result, theposition of the peak in the pattern of vibration varies depending on thefrequency of stimulation The cochlea and its components work thus as afrequency-to-position translation device Where the basilar membrane isdeflected most, there fire the hair cells of the organ of Corti; these interfacethe afferent neurones that carry signals to the higher levels of the auditorysystem The signals leaving the ear are therefore in the form of a frequencydomain representation The intensity of each frequency range (the exactnature and extent of these ranges is considered later) is coded by means of

a pulse rate modulation scheme

18 Newnes Guide to Digital TV

Figure 2.7 The physiology of the ear

whisper we can hear is a billionthð1012Þ of the intensity of the sound of a jetaircraft taking off heard at close range In engineering terms, you could sayhuman audition is equivalent to a true 20-bit system – 16 times better thanthe signal processing inside a compact disc player! Interestingly, the tiniestsound we can hear occurs when our eardrums move less than the diameter

of a single atom of hydrogen Any more sensitive, and we would be keptawake at night by the sound of the random movement of the nitrogenmolecules within the air around us In other words, the dynamic range ofhearing is so wide as to be up against fundamental physical limitations

Masking

The cochlea and its components work as a frequency-to-position translationdevice, the position of the peak in the pattern of vibration on the basilarmembrane depending on the frequency of stimulation Because of this itgoes without saying that the position of this deflection cannot be vanish-ingly small – it has to have some dimension This might lead us to expect thatthere must be a degree of uncertainty in pitch perception and indeed there

is, although it’s very small indeed, especially at low frequencies This isbecause the afferent neurones, which carry signals to the higher levels of theauditory system, ‘lock on’ and fire together at a particular point in thedeflection cycle (the peak) In other words, a phase-detection frequencydiscriminator is at work This is a truly wonderful system, but it has onedrawback; due to the phase-locking effect, louder signals will predominateover smaller ones, masking a quieter sound in the same frequency range.(Exactly the same thing happens in FM radio, where this phenomenon is

Table 2.1

street

known as capture effect.) The range of frequencies over which one soundcan mask another is known as a critical band, a concept due to Fletcher(quoted in Moore, 1989) Masking is very familiar to us in our daily lives Forinstance, it accounts for why we cannot hear someone whisper whensomeone else is shouting The masking effect of a pure tone gives us aclearer idea about what’s going on Figure 2.8 illustrates the unusual curvethat delineates the masking level in the presence of an 85 dBSPL tone Allsounds underneath the curve are effectively inaudible when the tone ispresent! Notice that a loud, pure sound only masks a quieter one when thelouder sound is lower in frequency than the quieter, and only then whenboth signals are relatively close in frequency Wideband sounds have acorrespondingly wide masking effect This too is illustrated in Figure 2.8,where you’ll notice the lower curve indicates the room noise in dBSPL inrelation to frequency for an average room-noise figure of 45 dBSPL (Noticethat the noise level is predominantly low frequency, a sign that the majority

of the noise in modern life is mechanical in origin.) The nearly parallel lineabove this room-noise curve indicates the masking threshold Essentiallythis illustrates the intensity level, in dBSPL, to which a tone of the indicatedfrequency would need to be raised in order to become audible Thephenomenon of masking is important to digital audio compression, as weshall see in Chapter 7

amplitude will mask sounds immediately preceding or following it in time;

as illustrated in Figure 2.9 When sound is masked by a subsequent signalthe phenomenon is known as backward masking, and typical quotedfigures for masking are in the range of 5–50 ms The masking effect thatfollows a sound is referred to as forward masking and may last as long 50–

200 ms, depending on the level of the masker and the masked stimulus

Unfortunately, the real situation with temporal masking is morecomplicated, and a review of the psychological literature reveals thatexperiments to investigate backward masking in particular dependstrongly on how much practice the subjects have received – with highlypractised subjects showing little or no backward masking (Moore, 1989).Forward masking is, however, well defined (although the nature of theunderlying process is still not understood), and can be substantial evenwith highly practised subjects

Film and television

Due to the persistence of vision, if the eye is presented with a succession

of still images at a sufficiently rapid rate, each frame differing only in thepositions moving within a fixed frame of reference, the impression isgained of a moving image In a film projector each still frame of film isdrawn into position in front of an intense light source whilst the source oflight is shut-off by means of a rotating shutter Once the film frame hasstabilized, the light is allowed through – by opening the shutter – and theimage on the frame is projected upon a screen by way of an arrangement

of lenses Experiments soon established that a presentation rate of about

12 still frames per second was sufficiently rapid to give a good impression

of continuously flowing movement, but interrupting the light source atFigure 2.9 The phenomenon of temporal masking

this rate caused unbearable flicker This flicker phenomenon was alsodiscovered to be related to the level of illumination; the brighter the lightbeing repetitively interrupted, the worse the flicker Abetted by the lowlight output from early projectors, this led to the first film frame-ratestandard of 16 frames per second (fps); a standard well above thatrequired simply to give the impression of movement and sufficientlyrapid to ensure flicker was reduced to a tolerable level when used withearly projection lamps As these lamps improved flicker became more of aproblem, until an ingenious alteration to the projector fixed the problem.The solution involved a modification to the rotating shutter so that, oncethe film frame was drawn into position, the shutter opened, then closed,then opened again, before closing a second time for the next film frame to

be drawn into position In other words, the light interruption frequencywas raised to twice that of the frame rate When the film frame rate waseventually raised to the 24 fps standard that is still in force to this day, thelight interruption frequency was raised to 48 times per second – a rate thatenables high levels of illumination to be employed without causingflicker

Television

To every engineer, the cathode ray tube (CRT) will be familiar enoughfrom the oscilloscope The evacuated glass envelope contains an elec-trode assembly and its terminations at its base, whose purpose it is toshoot a beam of electrons at the luminescent screen at the other end of thetube This luminescent screen fluoresces to produce light wheneverelectrons hit it In an oscilloscope, the deflection of this beam is effected

by means of electric fields – a so-called electrostatic tube In television, theelectron beam (or beams in the case of colour) is deflected by means ofmagnetic fields caused by currents flowing in deflection coils woundaround the neck of the tube where the base section meets the flare Such atube is known as an electromagnetic type

Just like an oscilloscope, without any scanning currents the televisiontube produces a small spot of light in the middle of the screen This spot oflight can be made to move anywhere on the screen very quickly by theapplication of the appropriate current in the deflection coils The bright-ness of the spot can be controlled with equal rapidity by altering the rate atwhich electrons are emitted from the cathode of the electron gunassembly This is usually effectuated by controlling the potential betweenthe grid and the cathode electrodes of the gun Just as in an electron tube

or valve, as the grid electrode is made more negative in relation to thecathode, the flow of electrons to the anode is decreased In the case of theCRT, the anode is formed by a metal coating on the inside of the tubeflare A decrease in grid voltage – and thus anode current – results in a

22 Newnes Guide to Digital TV